How to stress-test your plan against a 2008 crash

Last reviewed July 2026 · 5 min read

Average returns are kind to retirement plans; sequences are not. Two retirees can earn the same average over thirty years and end up in different worlds, because the one who eats a crash in year one sells depressed assets to fund spending and never gets that principal back. 2008 is the canonical test case — and Deorbit Plan can inject the real 2008 into your simulation, at the worst possible moment, to see whether your plan survives it.

What actually happened in 2008

In calendar 2008, US stocks (S&P 500, dividends included) returned about −36.6%, while 10-year Treasuries returned +20.1%, T-bills +1.4%, and inflation ran 3.8%. The year after, 2009, stocks rebounded +25.9% and those same Treasuries gave back −11.1%. Two lessons hide in those numbers: high-quality bonds were the only thing that worked during the crash, and whether you count the rebound changes the damage a lot — which is why the simulator makes the recovery a separate choice.

Step 1 — turn the stress test on

Open the Market Assumptions panel and find the Stress test group. Switch the selector from Off (no crash injected) to Replay the 2008 crash in one year. This replaces one simulated year’s market returns on every Monte Carlo path with 2008’s actual numbers — stocks, bonds, cash, and inflation together. It works under both return models (parametric and historical bootstrap) and leaves every other year’s draws untouched, so you are measuring one bad year, not a pessimistic re-tuning of the whole future.

Step 2 — choose when it hits

The “When does the crash hit?” selector offers three timings. First year of retirement is the classic worst case: the crash lands the same year on every path, right as withdrawals begin — pure sequence-of-returns risk, the scenario 2008’s actual retirees lived. One random year per path scatters the crash uniformly between retirement and the end of the plan (seeded, so runs are reproducible) — useful when you want “a 2008 happens sometime” rather than “at the worst time.” A specific calendar year lets you probe a particular fear, like the year a pension starts or the year before Medicare.

Step 3 — decide whether 2009 follows

The checkbox “Also replay the 2009 recovery the following year” forces the next year to 2009’s actual returns. Leaving it off is the harsher test: the year after the crash is drawn normally, so some paths crash and then merely muddle along — no guaranteed V-shaped rebound. Run it both ways; the gap between them is roughly the value of the recovery you are implicitly counting on.

Step 4 — read the damage

Compare the success gauge before and after — success means the money lasts to the horizon age, so the drop is the headline cost of the crash. On the Net worth across paths fan chart, watch the lower band: a plan whose 5th percentile stays above zero after the stress is genuinely robust. Turn on the Events toggle and a “2008-style market crash (stress test)” marker appears in the lane under each chart at the crash year (with random-per-path timing, the assumptions footnote notes that instead). The footnote also records the stress settings, so an exported or screenshotted result stays honest.

Step 5 — compare fairly, then shop for defenses

To A/B a defense, first set a fixed RNG seed in Market Assumptions so both scenarios face identical market histories, then load a variant as Scenario B and read the deltas in the compare view. The classic defense to test is the post-retirement bond tent in the Allocation group: arrive at retirement more conservative, then glide the stock share back up over a few years so the dangerous early window is defended. A stressed run with and without the tent shows exactly what that insurance buys — and what it costs in the paths where no crash comes early.

Try it in Deorbit Plan

In the Market Assumptions panel, set Stress test to Replay the 2008 crash in one year, choose First year of retirement, and toggle the 2009 recovery checkbox both ways. Watch the success gauge and the fan chart’s lower band, enable the Events marker lane, then use a fixed seed plus Scenario B to test the post-retirement bond tent against the same crash.

Educational content only — not financial, tax, or investment advice.

See how this plays out with your own numbers. Try it in the simulator →